Lipopolysaccharide Simulations Are Sensitive to Phosphate Charge and Ion Parameterization

Current simulations of bacterial outer membranes often rely on force field parameterizations that differ in phosphate protonation states and ion modeling. While prior research has shown that LPS bilayers form a key barrier against antibiotics, uncertainty remains about how protonation and ion parameters influence simulation accuracy. No prior work had resolved how these factors affect bilayer properties like hydrogen bonding and lipid spacing. This gap motivated a study to compare different parameter sets and their effects on LPS behavior. Existing models assume deprotonated phosphate groups at physiological pH, but this assumption may not align with experimental evidence. The need to reconcile computational and experimental data drove the investigation. Researchers aimed to clarify how protonation and ion selection impact bilayer stability and function. This work addresses a critical need in antibiotic development and membrane modeling.

The study aimed to determine how phosphate protonation and ion parameterization affect LPS bilayer simulations. Researchers focused on resolving discrepancies between computational models and experimental observations of LPS behavior. They sought to identify which parameter choices best reproduce known bilayer properties. The goal was to improve the accuracy of molecular dynamics simulations for LPS membranes. By comparing protonated and deprotonated LPS models, the team aimed to clarify which state dominates at physiological pH. They also tested how different cation types and parameter sets influence hydrogen bonding and lipid spacing. This work was intended to guide future simulations of bacterial membranes and drug interactions. The findings could help refine force fields for more reliable predictions of membrane behavior.

The researchers used molecular dynamics simulations to model LPS bilayers with various parameter sets. They tested different protonation states of the phosphate groups in lipid A. Each simulation included multiple LPS chemotypes and ion parameterizations. They compared monovalent and divalent cations with modified nonbonded parameters. Area per lipid and hydrogen bonding were measured to assess bilayer properties. Alchemical free energy simulations were used to estimate pKa values for LPS. These theoretical results were validated using 31P solid-state NMR experiments. The combination of computational and experimental approaches allowed direct comparison of protonation states.

Simulations showed that bilayer properties are highly sensitive to phosphate charge and ion parameters. Protonated LPS with monovalent cations matched experimental data best. Deprotonated models predicted different hydrogen bonding and lipid spacing. Area per lipid varied significantly depending on protonation state and cation type. Alchemical simulations predicted pKa values for LPS phosphate groups. These values were confirmed by 31P NMR experiments at physiological pH. The results indicated that protonated LPS dominates under normal conditions. This contradicts many existing force fields that assume deprotonated phosphate groups. The study also showed that ion parameterization strongly affects bilayer stability and function.

The study concludes that LPS simulations are highly sensitive to phosphate protonation and ion parameterization. Protonated LPS models better align with experimental evidence at physiological pH. Current force fields often assume deprotonated phosphate groups, which may lead to inaccurate predictions. The findings suggest that existing parameter sets should be updated for consistency. Researchers recommend using protonated LPS and monovalent cations in simulations. This approach improves agreement with experimental data on bilayer properties. The work highlights the importance of matching computational models to real-world observations. These conclusions support more accurate modeling of bacterial membranes and drug interactions.